UNESCO Recommendation on the Ethics of AI PA2b: paras 56, 58, 69 Strategies, certification, AI ethics officers and regulatory sandboxes

States are invited to adopt AI strategies at national and regional level and to weigh soft governance such as certification of AI systems with mutual recognition, scaled to the sensitivity of the domain, with audit levels, ongoing monitoring and re-certification, without excessive burden on small firms, civil society or research. Public bodies, companies and civil society should consider an independent AI Ethics Officer or similar mechanism to oversee impact assessment, auditing and monitoring, and a network of such officers is encouraged. Norms should be developed with all AI actors, and policy prototypes and regulatory sandboxes used to test laws before adoption, with support for local governments.